Format-Preserving Encryption for Databricks

What It Does

Cyphera for Databricks adds format-preserving encryption to your Spark environment. Upload the Cyphera JAR as a cluster library or to a Unity Catalog Volume, register the UDFs, and call cyphera_protect to encrypt and cyphera_access to decrypt from SQL or DataFrame code. Data Protection Headers are embedded in the output so access needs no configuration name. Encrypted values keep their original format — SSNs stay SSNs, phone numbers stay phone numbers.

Quick Example

SQL

SELECT cyphera_protect('ssn', '123-45-6789');
-- → 'T01948-37-2150' (DPH-formatted, format preserved)

SELECT cyphera_access(cyphera_protect('ssn', '123-45-6789'));
-- → '123-45-6789'

Register UDFs in a notebook

# Python
spark._jvm.io.cyphera.databricks.CypheraRegistrar.registerAll(spark._jsparkSession)

// Scala
io.cyphera.databricks.CypheraRegistrar.registerAll(spark)

How It Works

Build the JAR with Maven and deploy it to your Databricks workspace. You have two options: upload as a cluster library for automatic loading, or place it in a Unity Catalog Volume for more controlled access. Once the JAR is available, register the UDFs from any notebook and they become available across your SQL and DataFrame workloads.